Nvidia's Huang Tells G20 Not to Regulate AI 'Theoretically'

What happened

Nvidia CEO Jensen Huang used a G20 platform to warn policymakers against writing artificial intelligence rules aimed at speculative or "theoretical" harms, arguing that overly cautious regulation risks slowing a technology he says is already delivering measurable economic and scientific benefits 1. His comments landed amid a broader wave of AI announcements that illustrate exactly the kind of momentum he's pointing to: new regional R&D hubs, state-level investment pushes, university supercomputing projects, and, in at least one case, a local government pulling back on classroom AI use rather than embracing it 2345.

Nokia's move to open an AI-focused network automation research center in Riyadh is a case study in the commercial buildout Huang is defending. The facility is designed to build software letting telecom networks self-configure, self-repair, and cut energy use with minimal human oversight 2. That kind of infrastructure investment mirrors a pattern playing out across the U.S. as well, where officials in Wyoming and New York are framing AI adoption as an economic imperative rather than a risk to be managed cautiously.

In Wyoming, an op-ed from a state official argued that the state must position itself as a leader in AI rather than a bystander, framing the technology as already reshaping work, education, and research and urging that graduates be prepared for an AI-driven job market 3. In Buffalo, the newly launched Empire AI Beta supercomputer at the University at Buffalo is being promoted as the most powerful academic AI research machine in the country, a flagship example of publicly backed AI infrastructure meant to keep pace with private-sector capacity 4.

Against that backdrop of expansion, New York City under Mayor Mamdani has taken the opposite approach in one specific domain: a one-year ban on AI tools for most public school students, a restriction that stands out against the accelerationist tone coming from industry and other state governments 5.

Where the reporting agrees

Across these accounts, there is a shared premise that AI has moved past the experimental stage and is now a live factor in economic policy, infrastructure planning, and education 1234. Multiple sources describe institutions — corporate, state, and academic — racing to build capacity, whether through Nvidia's hardware-driven ecosystem, Nokia's regional R&D center, Wyoming's workforce planning, or Buffalo's supercomputing project 1234. There's also a consistent thread that governments are being pressed to make active choices about AI policy right now rather than waiting, whether that means Huang lobbying G20 officials against premature restrictions, Wyoming's push to lead, or Mamdani choosing to restrict rather than expand access 135.

Where it doesn't

The clearest divergence is in posture toward regulation and caution itself. Huang's message to the G20 explicitly frames concern over AI harms as "theoretical" and warns against rules built on speculative risk 1, while Mamdani's schools policy operates from the opposite assumption — that AI's effects on students are significant enough right now to justify a blanket one-year ban 5. These two accounts aren't describing the same policy domain, but they represent fundamentally different risk calculations happening in the same news cycle, one dismissing precaution and one actively practicing it.

There's also a difference in scale and specificity. Nvidia's and Nokia's stories involve concrete corporate infrastructure commitments with defined technical goals — self-healing telecom networks, hardware-scale compute 12 — while the Wyoming piece is more of an aspirational policy argument without comparable operational detail 3. The Buffalo item sits somewhere in between, naming a specific, ranked machine but relying on a single outlet's framing of its status as the country's most powerful academic AI research computer, a claim not corroborated elsewhere in this set 4.

The throughline

Taken together, the coverage does not resolve into a single verdict on AI regulation, and it shouldn't be forced into one. What the sources do support is that the debate Huang is trying to shape at the G20 is already being litigated in practice, not theory, at the local and institutional level — through infrastructure builds in Riyadh and Buffalo, workforce strategy in Wyoming, and a precautionary ban in New York City classrooms. The disagreement isn't really among the outlets; it's among the policymakers and executives they're covering, and that gap between industry's push for permissive rules and at least one government's move toward restriction is the story worth watching.